Re: USB enclosures seem to require read(16) with >2TB drives

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 9 Nov 2012, Jason J. Herne wrote:

> Hello, I've noticed Linux seems to have issues with external USB
> enclosures containing drives > 2 TB.   The USB mass storage driver
> apparently emulates a SCSI device and sends a read/write(10) for all

Actually, it's the SCSI disk driver which decides whether to use the 
10-byte or 16-byte commands.

> requests where the target sector is not large enough to require the
> use of read(16).  The issue is that both of the enclosures I was able
> to test with do not properly accept this read(10) command when a 3TB
> drive (Seagate ST3000DM) is installed.  A wireshark USB trace shows
> the read(10) command is sent to the drive and the drive never
> responds.  The same drive/enclosure used in Windows 7 shows that
> read(16) is always used to communicate with the drive and the drive
> works without issue with Windows 7.
> 
> A few notable points:
> 1. The drive operates without issue when connected to an internal Sata
> port on my computer.
> 2. The enclosure operates without issue with Linux (using read(10) )
> when a <2TB drive is installed.
> 3. I have reproduced this problem on several computers, and using
> several Linux kernels, including 3.7.0-rc4-next.
> 
> I first reported this to the linux-usb list thinking it was a USB
> issue.  At the request of Alan Stern I am also reporting it here.
> 
> I'm not sure if I am taking the correct approach but I have written a
> simple patch that resolves the problem for me.  In essence, the patch
> adds a new bit to scsi_device that indicates it should favor
> read/write(16) over read/write(10).  sd_prep_fn is modified to check
> this bit, and use a 16 byte operations when a large capacity drive is
> in use and this bit is on. Finally, the USB storage driver's scsi glue
> code is modified to turn this bit on for all of its devices.  Not sure
> if this is the correct fix, but it works for me.  I'm willing to

It's not correct, because it will break USB drives that don't recognize
the 16-byte commands.

The bit should be set only for disks that have 2^32 or more blocks, or
maybe only for disks that hold 2^41 or more bytes (and maybe only for
USB disks).  Since usb-storage doesn't know the disk's capacity, the
decision about whether to set the new bit has to be made in the SCSI
disk driver.

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [SCSI Target Devel]     [Linux SCSI Target Infrastructure]     [Kernel Newbies]     [IDE]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux ATA RAID]     [Linux IIO]     [Samba]     [Device Mapper]
  Powered by Linux